Communication pubs.acs.org/JACS
Chloride Oxidation by Ruthenium Excited-States in Solution Sara A. M. Wehlin, Ludovic Troian-Gautier, Guocan Li, and Gerald J. Meyer* Department of Chemistry, University of North Carolina at Chapel Hill, Murray Hall 2202B, Chapel Hill, North Carolina 27599-3290, United States S Supporting Information *
ABSTRACT: Photodriven HCl splitting to produce solar fuels is an important goal that requires strong photooxidants capable of chloride oxidation. In a molecular approach toward this goal, three ruthenium compounds with 2,2′-bipyrazine backbones were found to oxidize chloride ions in acetone solution. Nanosecond transient absorption measurements provide compelling evidence for excited-state electron transfer from chloride to the Ru metal center with rate constants in excess of 1010 M−1 s−1. The Cl atom product was trapped with an olefin. This reactivity was promoted through pre-organization of ground-state precursors in ion pairs. Chloride oxidation with a tetra-cationic ruthenium complex was most favorable, as the dicationic complexes were susceptible to photochemical ligand loss. Marcus analysis afforded an estimate of the chlorine formal reduction potential E°(Cl•/−) = 1.87 V vs NHE that is at least 300 meV more favorable than the accepted values in water.
Figure 1. [Ru(bpz)3]2+ (1), [Ru(bpz)2(tmam)]4+ (2), and [Ru(bpz)2(deeb)]2+ (3), where bpz is 2,2′-bipyrazine, tmam is 4,4′trimethylamino-methyl-2,2′-bipyridine, and deeb is 4,4′-diethylester2,2′-bipyridine.
shown that [Ru(bpz)2(deeb)]2+ can oxidize bromide to bromine in acetone.12 The tmam ligand, on the other hand, possesses a dicationic charge and a halide binding site, as has previously been shown with iodide.9 Our approach was to combine the strong oxidizing power conferred by bpz ligands with the ion-pairing ability of the tmam ligand in complex 2, to pre-organize the ground-state reactants in close proximity prior to light absorption so as to promote reactivity. The complexes displayed photophysical properties common to most ruthenium polypyridyl species, such as ground-state absorption features around 450 nm, assigned to metal-to-ligand charge-transfer (MLCT) transitions, and room-temperature photoluminescence (Figure 2a) that decayed exponentially to the ground state.11 Photophysical properties of the complexes are summarized in Table 1. The values obtained for 1 and 3 in acetone were in good agreement with previously published data acquired in acetonitrile.7,8,13,14 Ion-pairing between 2 and chloride was investigated in titration experiments monitored by 1H NMR and UV−vis spectroscopy. Additions of TBACl induced shifts that increased with the chloride concentration in several 1H NMR resonances (Figure S3). A large, downfield shift (Δppm > 1) was observed for the 3,3′ protons and the methylene protons on the tmam ligand, the latter of which resolved into a roofed doublet upon chloride addition due to their enantiotopic nature, indicating that the tmam ligand was an ion-pairing binding site. It is worth noting that the typical H-3,3′ binding sites on the bpz ligands were unaffected,12 but spectral shifts associated with the 5,5′ and 6,6′ H atoms were observed, indicating a second binding site near the metal center. The equilibrium constant was extracted from a modified Benisi−Hildebrand analysis to be 44 000 ± 5000 M−1.15,16 The broad resonances indicated a dynamic process on the NMR time scale.
B
ond formation as a result of an electron-transfer reaction from an excited state is of great interest for solar energy conversion and storage, especially when the excited states are generated by visible-light absorption. A specific kind of reaction that has gained more attention recently is HX splitting, where X is a halide, which produces H2 and X2.1,2 When moving up the halide group, reduction potentials become progressively more positive,3,4 and more potent photo-oxidants are required. Nevertheless, chloride is a very promising target, as it is a naturally abundant resource, making up more than 3% by weight of seawater, which in turn corresponds to 97% of our water supply.5 Furthermore, chlorine is used in a number of industries and is produced in the chlor-alkali process, which employs electrolysis to produce chlorine from sodium chloride.6 Herein we report on the photophysical and photochemical behavior of three ruthenium complexes, [Ru(bpz) 3 ] 2+ (1), [Ru(bpz) 2 (tmam)] 4 + (2), and [Ru(bpz)2(deeb)]2+ (3) (Figure 1), and their use for successful chloride oxidation with visible light. Complexes 1 and 3 have been previously reported,7,8 whereas complex 2 was synthesized by reaction between [Ru(bpz)2Cl2] and [tmam]Cl2 (see Supporting Information (SI)).8,9 The complexes were characterized by 1H NMR and mass spectrometry (SI). All experiments were performed in acetone at room temperature, unless otherwise stated. It is generally accepted that ruthenium complexes bearing at least two 2,2′-bipyrazine (bpz) ligands behave as potent excited-state oxidants.7,10,11 Furthermore, it was recently © 2017 American Chemical Society
Received: June 29, 2017 Published: August 30, 2017 12903
DOI: 10.1021/jacs.7b06762 J. Am. Chem. Soc. 2017, 139, 12903−12906
Communication
Journal of the American Chemical Society
Figure 2. (a) Ground-state absorption and photoluminescence spectra of [Ru(bpz)3]2+, [Ru(bpz)2(tmam)]4+, and [Ru(bpz)2(deeb)]2+ in acetone. (b) Absorption of [Ru(bpz)2(tmam)]4+ upon the addition of TBACl in acetone; the inset recasts the same data as difference spectra with the spectrum at 0 equiv of chloride as a reference. (c) Time-resolved and steady-state (inset) photoluminescence of [Ru(bpz)2(tmam)]4+ with added chloride. (d) Stern−Volmer plots for excited-state chloride quenching. (e) Transient absorption spectra measured at the indicated delay times after pulsed 450 nm excitation of [Ru(bpz)2(tmam)]4+ with 2 equiv of chloride in acetone. (f) Transient absorption spectra for the monoreduced and excited states of [Ru(bpz)2(tmam)]4+ after pulsed 450 nm excitation (scatter). Overlaid (solid lines) are the decay-associated spectra.
agreement with steady-state photolysis experiments that revealed the same order for ligand loss photochemistry (Figure S6). Electrochemical data measured with 2 are summarized in Table 2, along with literature data for 1 and 3. The similarity of
Table 1. Photophysical Properties of Complexes 1, 2, and 3 in Acetone
a
Ru
Abs (nm) (ε, M−1 cm−1))
PL λmax (nm)
τ (μs)
ΦPL
kr (×104 s−1)
knr (×105 s−1)
1 2 3a
440 (13 000) 450 (12 600) 450 (14 200)
610 635 630
0.90 1.50 1.75
0.060 0.038 0.090
6.0 2.5 5.1
9.4 6.4 5.2
Table 2. Formal Reduction Potentials and Excited-State Free Energiesa
Values from ref 12.
E° (V)
Visible absorption spectra further confirmed ion-pair formation between complex 2 and chloride in solution, as the MLCT absorption band decreased in intensity and red-shifted (Figure 2b) without evidence for ligand photosubstitution chemistry. At high chloride concentrations a precipitate formed. The absorption spectra of complexes 1 and 3 showed no indications of ion-pair formation. With complex 1, a precipitate formed in the presence of chloride, indicative of the formation of a less soluble chloride salt. With the addition of approximately 10% water, the complex became soluble, further confirming the formation of the chloride salt (Figure S5). In a 9:1 acetone/water solution complex 1 underwent rapid ligand substitution when excited by 460 nm light (Figure S6). Complex 3 underwent rapid ligand exchange with chloride, even under ambient light conditions. Such photochemistry is known and has been attributed to the population of the metalcentered (MC) states, sometimes called ligand-field (LF) states. These states are antibonding in character and therefore often lead to ligand loss chemistry. The activation energy for 3MLCT → MC internal conversion was quantified for the three complexes through temperature dependent lifetime measurements with Arrhenius-type analysis.17 Butyronitrile was utilized as a solvent as it provided a wider temperature window than acetone.18 The Ea values were 2700, 3100, and 2600 cm−1 for complexes 1, 2, and 3, respectively. This trend was in good
complex 1 2 3
RuIII/II c
2.23 2.10 2.04d
bpz−/0
(Run+*/(n−1)+)
ΔGES (eV)b
−0.43 −0.50 −0.48d
1.8 1.7 1.7
2.26 2.18 2.18e
c
a
Unless otherwise noted all values were measured in acetonitrile at room temperature and were corrected to NHE. bMeasured in acetone. c Data from ref 7. dData from ref 13. eData from ref 12.
the first reduction potential indicated that the bpz ligand was first reduced for all the complexes, and hence the excited state was similarly localized on a bpz ligand.19 The first reduction of 2 was reversible, but at more negative potentials discoloration was observed on the electrode, similar behavior has been reported for related cationic complexes.20 Metal based RuIII/II potentials were determined at highly positive potentials. The oxidizing power of the MLCT excited states, E°(Run+*/(n−1)+), where n is the charge of the complex cation, were determined from ground-state potentials and the free energy stored in the excited state, ΔGES. All complexes were potent excited-state oxidants with reduction potentials ranging from 1.7 to 1.8 V vs NHE. Complex 1 was the strongest excited-state oxidant due to the three 2,2′-bipyrazine ligands. Significant excited-state quenching by chloride was evident for all three complexes (Figure 2c,d and Table 3). Stern Volmer 12904
DOI: 10.1021/jacs.7b06762 J. Am. Chem. Soc. 2017, 139, 12903−12906
Communication
Journal of the American Chemical Society
the exact nature of the photochemical product(s) is unknown (Figure S10). The resonances associated with DMBD decreased by a factor of 2, while the TBA+ resonances remained unchanged, providing compelling evidence for the production of Cl atoms in the excited-state reaction. There exists a substantial literature of aqueous halide redox potentials that have been predominantly determined from pulse radiolysis data; however, values in organic solvents are less common.3 Furthermore, there is significant variance in the aqueous Cl•/− reduction potentials available in the literature, ranging from 2.2 to 2.4 V.3 An estimate of the chloride reduction potential in acetone was calculated using Marcus theory. The observed dynamic quenching rate constant kq, for complex 2, was corrected for diffusion and excited-state encounter complex formation (SI).27,28 Assuming reorganization energy of 1 eV, the free energy change for the excited-state electron-transfer reaction, ΔG°, was estimated from Marcus theory to be +0.19 eV. This ΔG° value and the E°(Ru4+*/3+) potential of 2* provided a formal reduction potential of E°(Cl•/−) = 1.87 V vs NHE for chlorine in acetone. This Marcus analysis indicates that chloride oxidation is thermodynamically uphill for 2*, although it occurs with a large rate constant. This is reminiscent of prior reports of rapid iodide oxidation by MLCT excited states when ΔG° ≈ 0, behavior attributed to the formation of inner-sphere adducts that enhances the electronic coupling prior to electron transfer.27 Thus, the possibility of an inner-sphere electrontransfer mechanism was ignored in this calculation that could account for the rapid reactivity.29−32 Nevertheless, the estimated Cl•/− potential is significantly less positive than the accepted values in water.3 Hence, acetone plays two important roles in this photoredox chemistry: it affects the Cl•/− reduction potential and provides a low dielectric continuum that promotes ion-pair formation that was particularly important for complex 2. In conclusion, the three complexes [Ru(bpz)3]2+ (1), [Ru(bpz)2(tmam)]4+ (2), and [Ru(bpz)2(deeb)]2+ (3) were found to photo-oxidize chloride in acetone or acetone/water solutions with rate constants greater than 1010 M−1 s−1. Nanosecond transient absorption spectroscopy revealed a mechanism wherein chloride transfers an electron to the excited state to yield the reduced ruthenium complex. Studies with radical traps indicated that the Cl atom was the other product. Complex 1 was the most potent photo-oxidant; however, the combination of oxidizing power conferred by 2,2′bipyrazine ligands with the ion-pairing ability of a dicationic tmam ligand induces greater stability in the ground and excited states of 2. The activation energy for internal conversion from the 3MLCT to MC states was quantified for all three complexes and was correlated with ligand loss photochemistry with chloride. The data show that appropriate ligand design and use of non-aqueous solvents can enable rapid chloride oxidation by molecular excited states.
Table 3. Stern−Volmer Parameters for the Complexes in Solution with TBACl KSV (M−1)
a
complex
I0/I
τ0/τ
kq (1010 M−1 s−1)
1 2 3a
64 000 − 27 000
61 000 19 000 32 000
7.0 ± 0.7 1.2 ± 0.1 ∼2
Measured in acetone/water (9:1).
analysis of lifetime and intensity data for 1* and 3* were linear with similar slopes, consistent with a dynamic electron-transfer mechanism (Figure 2d). Note that the data for 3* were obtained in a 9:1 acetone/water mixture in an effort to minimize ligand loss photochemistry and to maximize solubility. Stern−Volmer analysis of 2* was complicated by the appearance of non-single exponential kinetics. The data were well described by a biexponential model with a short lifetime of 75 ns that was independent of the chloride concentration and a longer lived excited state that was dynamically quenched by chloride.12 Quenching of excited states by chloride has been previously reported, but the mechanism(s) remain speculative.14,16,18,21,22 In an effort to unambiguously identify the quenching mechanism herein, transient absorption spectroscopic measurements were performed. In neat acetone, pulsed 450 nm excitation of 2 showed spectra consistent with the formation of a MLCT excited state, with ground-to-excited-state isosbestic points at 400 and 500 nm. Upon addition of 2 equiv of chloride, a new long-lived feature was observed with a maximum at 510 nm (Figure 2e). This was assigned to the one-electron-reduced form of 2, whose spectrum was independently measured (Figure S7) by published procedures.23 A similar feature with a peak at 510 nm was observed for complex 1 (Figure S8). In contrast, no redox chemistry was observed when weaker photo-oxidants were employed. For example, a related complex, [Ru(bpy)2(tmam)]4+* showed no evidence for excited-state quenching under the same conditions (Figure S9). Hence the observation of the one-electronreduced ruthenium complex after light excitation of 1 or 2 indicated reductive electron transfer from chloride to the excited complex. Excited-state electron transfer from Cl− is expected to form the chlorine atom (Cl•). Pulse radiolysis data in aqueous solutions indicate that Cl• reacts with Cl− to form the radical anion (Cl2•−), and that both Cl• and Cl2•− absorb light in the 340−360 nm region.24,25 Transient absorption experiments in this wavelength region, performed in an attempt to identify the Cl− photo-oxidation products, showed a very weak absorption on time scales that precluded excited-state participation but could not be conclusively assigned to oxidized chloride species. Since the transient data did not unambiguously identify the chloride oxidation product(s), steady-state 460 nm (4.5 mW/ cm2) illumination of 2 was carried out in the presence of a halogen trap, 2,3-dimethyl-1,3-butadiene (DMBD), in deuterated acetone.26 The 1H NMR analysis revealed that net photochemistry had occurred. Control experiments indicated that all three components, Cl − , DMBD, and [Ru(bpz)2(tmam)]4+, were required for this photochemical reaction to occur. Chlorine atom addition to an alkene was expected to result in downfield shifts due to the electronegativity of the halide, which was indeed observed although
■
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.7b06762. Experimental details, characterization for newly reported compounds, 1H NMR titration, and transient absorption spectroscopy data, including Figures S1−S10 (PDF) 12905
DOI: 10.1021/jacs.7b06762 J. Am. Chem. Soc. 2017, 139, 12903−12906
Communication
Journal of the American Chemical Society
■
(27) Farnum, B. H.; Gardner, J. M.; Meyer, G. J. Inorg. Chem. 2010, 49 (22), 10223−10225. (28) Sutin, N. Acc. Chem. Res. 1982, 15 (9), 275−282. (29) Yonemoto, E. H.; Riley, R. L.; Kim, Y. I; Atherton, S. J.; Schmehl, R. H.; Mallouk, T. E. J. Am. Chem. Soc. 1992, 114, 8081− 8087. (30) Troian-Gautier, L.; Beauvilliers, E. E.; Swords, W. B.; Meyer, G. J. J. Am. Chem. Soc. 2016, 138, 16815−16826. (31) Cooley, L. F.; Larson, S. L.; Elliott, C. M.; Kelley, D. F. J. Phys. Chem. 1991, 95, 10694−10700. (32) Roest, M. R.; Verhoeven, J. W.; Schuddeboom, W.; Warman, J. M.; Lawson, J. M.; Paddon-Row, M. N. J. Am. Chem. Soc. 1996, 118, 1762−1768.
AUTHOR INFORMATION
Corresponding Author
*
[email protected] ORCID
Ludovic Troian-Gautier: 0000-0002-7690-1361 Gerald J. Meyer: 0000-0002-4227-6393 Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We acknowledge support from the National Science Foundation (NSF), award CHE-1213357. S.A.M.W. thanks the Duke Energy Graduate Fellowship. L.T.-G. acknowledges the Belgian American Educational Foundation (BAEF) as well as the Bourse d’Excellence Wallonie-Bruxelles (WBI.World) for generous support. The authors would like to thank Dr. R. N. Sampaio for insightful discussions.
■
REFERENCES
(1) McDaniel, N. D.; Bernhard, S. Dalton Trans. 2010, 39, 10021. (2) Esswein, A. J.; Nocera, D. G. Chem. Rev. 2007, 107, 4022−4047. (3) Stanbury, D. M. Adv. Inorg. Chem. 1989, 33, 69−138. (4) Wardman, P. J. Phys. Chem. Ref. Data 1989, 18, 1637−1755. (5) Du, J.; Chen, Z.; Chen, C.; Meyer, T. J. J. Am. Chem. Soc. 2015, 137, 3193−3196. (6) Lakshmanan, S.; Murugesan, T. Clean Technol. Environ. Policy 2014, 16, 225−234. (7) Rillema, D. P.; Meyer, T. J.; Conrad, D.; Allen, G. Inorg. Chem. 1983, 22, 1617−1622. (8) Bergeron, B. V.; Meyer, G. J. J. Phys. Chem. B 2003, 107, 245− 254. (9) Swords, W. B.; Li, G.; Meyer, G. J. Inorg. Chem. 2015, 54, 4512− 4519. (10) Kajouj, S.; Marcélis, L.; Lemaur, V.; Beljonne, D.; Moucheron, C. Dalton Trans. 2017, 46, 6623−6633. (11) Troian-Gautier, L.; Moucheron, C. Molecules 2014, 19, 5028− 5087. (12) Li, G.; Ward, W. M.; Meyer, G. J. J. Am. Chem. Soc. 2015, 137, 8321−8323. (13) Farnum, B. H.; Jou, J. J.; Meyer, G. J. Proc. Natl. Acad. Sci. U. S. A. 2012, 109, 15628−15633. (14) Allen, G.; White, R.; Rillema, D. P.; Meyer, T. J. J. Am. Chem. Soc. 1984, 106, 2613−2620. (15) Benesi, H. A.; Hildebrand, J. H. J. Am. Chem. Soc. 1949, 71, 2703−2707. (16) Ward, W. M.; Farnum, B. H.; Siegler, M.; Meyer, G. J. J. Phys. Chem. A 2013, 117, 8883−8894. (17) Van Houten, J.; Watts, R. J. J. Am. Chem. Soc. 1976, 98, 4853− 4858. (18) Rebarz, M.; Marcélis, L.; Menand, M.; Cornut, D.; Moucheron, C.; Jabin, I.; Kirsch-De Mesmaeker, A. Inorg. Chem. 2014, 53, 2635− 2644. (19) Crutchley, R. J.; Lever, A. B. P. J. Am. Chem. Soc. 1980, 102, 7128−7129. (20) Zysman-Colman, E.; Slinker, J. D.; Parker, J. B.; Malliaras, G. G.; Bernhard, S. Chem. Mater. 2008, 20, 388−396. (21) Kalyanasundaram, K. J. Phys. Chem. 1986, 90, 2285−2287. (22) Masschelein, A.; Jacquet, L.; Kirsch-De Mesmaeker, A.; Nasielski, J. Inorg. Chem. 1990, 29, 855−860. (23) DiMarco, B. N.; O’Donnell, R. M.; Meyer, G. J. J. Phys. Chem. C 2015, 119, 21599−21604. (24) Nagarajan, V.; Fessenden, R. W. J. Phys. Chem. 1985, 89, 2330− 2335. (25) Treinin, A.; Hayon, E. J. Am. Chem. Soc. 1975, 97, 1716−1721. (26) Heyduk, A. F.; Nocera, D. G. Science 2001, 293, 1639−1641. 12906
DOI: 10.1021/jacs.7b06762 J. Am. Chem. Soc. 2017, 139, 12903−12906
